We aimed to assess the role of genomic HLA-I/II heterozygosity in the overall survival benefit in patients with unresectable locally advanced, metastatic non-small lung cancer treated by PD1/L1 inhibitors.
We collected blood from 170 advanced lung cancer patients treated with immunotherapy at two major oncology centres in Western Australia. High quality DNA was extracted from white blood cells and used for HLA-I/II typing. Information of tumour PDL1 status, pre-treatment neutrophil to lymphocyte ratio (NLR) and sex were identified. Each of those variables were correlated independently then in a multivariate analysis with OS. Correlation between HLA heterozygosity and response was assessed using Fisher exact. Patients with reduced or stable disease for 6 or more months were considered as responders.
Our data matured for 146 patients treated with anti-PD1/L1 therapy. We found that heterozygosity at all HLA-I loci had a favorable statistical trend towards better OS(HR = 0.5,P=0.06). However, heterozygosity at all HLA-II loci was not associated with OS(HR = 0.91,P=0.76). While response rate was higher amongst heterozygous patients especially at HLA-I (53.4% vs 45.6%), it was not statistically significant. NLR>5 was associated with statistically significant worse survival (HR = 2.22,P=0.009). This effect was driven by the patients with PDL1≥50% (HR = 3.86, P = 0.01 vs HR = 0.7, P = 0.5 for patients with PDL1<50%). Similarly, the effect of sex was seen mainly among patients with PDL1≥50%, with men more likely to have better OS than women (HR = 0.36, P = 0.06 vs HR = 1.1, P = 0.89 in the group of patients with PDL1<50%). A multivariate analysis showed that all three variables to be statistically significant predictors of survival. Table: LBA14Variable HR(95% CI) P value Sex 1.9(1.1-3.5) 0.028 NRLD(≥5vs<5) 7.1(1.8-27.3) 0.004 Homozygosity at ≥ 1 HLA-I 0.3(0.8-0.9) 0.039
Our analysis suggest that OS among advanced lung cancer patients treated with immunotherapy is more likely to be influenced by homozygosity at HLA-I loci. Women with PDL1≥50%, NLR>5 and homozygous at ≥ 1 HLA-I locus possibly carries the worse prognosis when treated with immunotherapy alone and might benefit from treatment escalation.
A/Prof Elin Gray, Edith Cowan University, Western Australia.
South Metropolitan Health Service - WA Health.
Fiona Stanley Hospital (FSH) and Sir Charles Gairdner Hospital (SCGH)
All authors have declared no conflicts of interest.
Individual patient data from two multicentre, randomised trials comparing atezolizumab with docetaxel in previously treated advanced non-small-cell lung cancer (NSCLC) were analyzed to identify any preferable therapeutic approaches for patient subsets stratified by EGFR mutation status.
We pooled 1,137 patients from the phase II POPLAR trial (NCT01903993) and the phase III OAK trial (NCT02008227). Patients were randomly assigned to receive either atezolizumab or docetaxel and were analyzed based on stratification of EGFR mutation status. We built a novel algorithm integrating the blood-based tumor mutation burden and the ctDNA maximum somatic allele frequency (bTMB-MSAF score) and developed a novel clinicopathologic-genomic nomogram to predict individual survival.
In the whole intention-to-treat population, OS was significantly longer with atezolizumab than with docetaxel (hazard ratio [HR] 0.72, P < 0.001). Among patients without EGFR or ALK mutations, overall survival (OS) was significantly longer with atezolizumab than with docetaxel (HR 0.67; P < 0.001); patients with a bTMB-MSAF score<20 showed improvement in OS (HR 0.56; P < 0.001) and progression-free survival (PFS) (HR 0.74; P = 0.0023), and these patients who concurrently had programmed death ligand 1 (PD-L1) expressing on over 50% tumor cells or over 10% of tumor-infiltrating immune cells (TC3 or IC3) had the greatest OS improvements (HR 0.44; P = 0.007) and promising PFS benefits (HR 0.54; P = 0.019). EGFR mutant patients could not gain significant OS or PFS benefits from atezolizumab over docetaxel, irrespective of PD-L1 expression and bTMB-MSAF score. Patients with low-risk scores had longer OS (HR 0.17; P < 0.001; AUC=0.912 for 3-year OS) than patients with high-risk scores classified by the nomogram.
Atezolizumab showed better OS and PFS than docetaxel in previously treated EGFR wild-type NSCLC patients with a bTMB-MSAF score < 20, especially in those with PD-L1 expression of TC3 or IC3. Our clinicopathologic-genomic nomogram is effective in predicting survival of NSCLC patients undergoing atezolizumab.
NCT01903993 and NCT02008227.
Herui Yao.
Herui Yao is funded by the National Science and Technology Major Project under Grant [2020ZX09201021]; National Natural Science Foundation of China under Grant [81372819, 81572596, U1601223]; Natural Science Foundation of Guangdong Province under Grant [2017A030313828]; and Guangzhou Science and Technology Program under Grant [201704020131].
All authors have declared no conflicts of interest.
A recent study has demonstrated that high blood tumor mutational burden (bTMB) was associated with significant improvements in PFS from atezolizumab in NSCLC. However, a prospective, phase II, B-F1RST study did not confirm the result. Multiple genetic mutations may result in resistance to therapy, including immunetherapy. Therefore, we speculated that low bTMB might be a favorable prognostic biomarker for immunetherapy and both high and low bTMB patients could derive benefit from atezolizumab. We thus investigated the non-linear association between bTMB and PFS, and tried to find new cut-off values.
This study used the clinical and bTMB data from POPLAR (n = 105, training set) and OAK (n = 324, validation set) studies. The non-linear association between bTMB and PFS was assessed using restricted cubic spline (RCS). The cut-off values for bTMB were calculated by using X-tile software.
In training set, bTMB showed an upside down J shape curve with PFS in RCS plot, suggesting a non-linear relationship between bTMB and PFS (P for non-linear < 0.001). The cut-off values of bTMB for predicting PFS were 7 and 14 mutations/Mb, and all patients were claasified into low (≤ 7 mutations/Mb), medium (8 ≤ bTMB ≤ 13 mutations/Mb), and high bTMB (≥ 14 mutations/Mb) groups according to the cut-off values. The median PFS and OS of patients with low and high bTMB were significantly longer than those of patients with medium bTMB in multivariate analysis. Similar results were shown in the validation set and the combined set (Table). Table: 318O Prognostic value of bTMB for PFS and OS in the training, validation, and combined data setsPFS OS HR 95% CI P value HR 95% CI P value Training data set (n = 105) Low bTMB 0.367 0.212-0.637 <0.001 0.441 0.239-0.814 0.009 Medium bTMB 1 Reference 1 Reference High bTMB 0.207 0.105-0.410 <0.001 0.336 0.161-0.704 0.004 Validation data set (n = 324) Low bTMB 0.689 0.502-0.945 0.021 0.644 0.441-0.939 0.022 Medium bTMB 1 Reference 1 Reference High bTMB 0.644 0.457-0.909 0.012 0.637 0.431-0.943 0.024 Combined data set (n = 429) Low bTMB 0.596 0.456-0.779 <0.001 0.649 0.476-0.884 0.006 Medium bTMB 1 Reference 1 Reference High bTMB 0.558 0.415-0.750 <0.001 0.689 0.496-0.958 0.027
There was a non-linear association between bTMB and survival in NSCLC patients receiving atezolizumab. Both high and low bTMB were associated with better clinical benefit with atezolizumab.
The authors.
Has not received any funding.
All authors have declared no conflicts of interest.
Immune checkpoint inhibitors (ICIs) therapy has been a pivotal treatment for NSCLC. However, additional biomarkers should be found out to cover more patients who may derive the greatest benefit from ICIs. Frameshift mutation by insertion or deletion (fsindel) has come to prominence due to its higher immunogenicity. Previous study has identified a positive correlation between fsindel and favorable clinical benefit in NSCLC. But it still requires validation. Therefore, we conducted a study to further assess the predictive role of fsindel.
A publicly available cohort of 385 ICIs-treated NSCLC patients from MSKCC were analyzed. We categorized patients into two groups; 0 fsindel (FS-) and more than 1 fsindel (FS+). The OS, PFS and response to ICIs therapy (ORR, DCR and DCB) were evaluated. We also developed a combined model of TMB and fsindel to further clarify the role of fsindel.
214 patients (55.58%) were found to be fsindel present (FS+). 189 patients (88.32%) were treated with PD-1/PD-L1 inhibitor monotherapy. Among the 356 patients with OS data, the OS of the FS+ patients was similar with that of FS- patients (median OS: 11 months vs. 12 months, P = 0.615). 240 patients (131 FS+, 109 FS-) were able to be evaluated for response to ICIs therapy. The median PFS was similar between FS+ and FS- group (median PFS, 3.43 months vs. 3.17 months, P = 0.114). The presence of fsindel was correlated with higher ORR (24.43% vs 15.60%, P = 0.091), DCR (56.49% vs 53.21%, P = 0.611) and DCB (35.77% vs. 24.04%, P = 0.055). TMB-H patients show significant difference in OS (median OS: 15 months vs. 10 months, P = 0.017) and marginally-significant difference in PFS (median PFS, 4.73 months vs. 2.60 months, P = 0.051). Moreover, TMB-H and FS+ patients had significantly better PFS and OS compared to patients who had either TMB-H or FS+ or neither (TMB-L and FS-) (median PFS = 6.17 months, P pfs = 0.004; median OS = 19 months, P os = 0.013).
Our study supported that fsindel might serve as a synergistic biomarker which could predict the efficacy of ICIs therapy for NSCLC patients. It might help us to precisely identify patients who could derive more benefit from ICIs therapy, especially in the fsindel present patients with higher TMB.
The authors.
National Key R&D Program of China (Grant No.2016YFC0905500, 2016YFC0905503).
All authors have declared no conflicts of interest.
Immune checkpoint blockade (ICB) therapy has been a pivotal treatment for lung cancer, yet predictive biomarkers are still lacking. PBRM1 mutation is associated with greater sensitivity to immunotherapy for clear cell renal cell carcinoma. To our knowledge, the frequency and clinical relevance of PBRM1 mutation in lung cancer remain unknown. Therefore, we conducted a retrospective study to evaluate the prevalence of PBRM1 mutation and its preliminary response to ICB therapy in NSCLC.
We analyzed the combined NSCLC cohort of 2767 patients, from 3 sources: (1) The Cancer Genome Atlas (TCGA) (N = 1144), (2) Memorial Sloan Kettering Cancer Center (MSKCC) (N = 1567), and (3) Dana Farber Cancer Institute (DFCI) (N = 56). We first estimated the prevalence of PBRM1 mutation in the whole NSCLC cohort. A subset of ICB-treated patients (N = 441) with annotated clinical records were further analyzed for association between PBRM1 mutation and response to ICB therapy. We also calculated the overall survival (OS) of 454 non-ICB treated patients. Institutional review board approval and informed consent were waived because all data were de-identified and publicly available.
Of 2767 patients included in our study, PBRM1 mutation was identified in 75 NSCLC patients (2.70%). Among 39 PBRM1-mutant patients with annotated clinical records, 25 patients (64.1%) were treated with PD-1/PD-L1 inhibitor monotherapy. In the cohort of ICB-treated patients (N = 441, PBRM1 MT=25), the OS of the PBRM1-mutant patients was worse than that of those without mutation (P = 0.03, median OS 6 vs. 13 months). In total, 14 patients, all with PBRM1 mutation, were able to be evaluated for response to ICB therapy. The median PFS was 2.1 months. The ORR was 28.6%, the DCR was 50%, and the DCB was 14.29%. In the cohort of non-ICB-treated patients (N = 454, PBRM1 MT=14), there seems to be no difference between the OS of the PBRM1 mutation subgroup and PBRM1 wild type subgroup (P = 0.097).
Our findings suggested that PBRM1-mutant NSCLC patients might get less survival benefit from ICB therapy, unlike previously reported data in clear cell renal cell carcinoma. Further prospective research is warranted to confirm the negative predictive role of PBRM1 in NSCLC ICB therapy.
The authors.
This work was supported by: National Key R&D Program of China (Grant No. 2016YFC0905500, 2016YFC0905503), Chinese National Natural Science Foundation project (Grant No. 81872499, 81772476), Science and Technology Program of Guangdong (Grant No. 2017B020227001), and Science and Technology Program of Guangzhou (Grant No. 201607020031, 201704020072).
All authors have declared no conflicts of interest.
The administration of immunotherapeutic antibodies against immune checkpoint proteins has shown great promises in the treatment of cancer patients. However, the response rate remains low, suggesting a strong need for predictive biomarkers. While PD-L1 is a commonly accepted biomarker, it is inadequate. Circulating microRNAs (miRNAs) have been shown as predictive biomarkers in various types of cancer therapies. We sought to determine whether miRNAs could predict response to immune checkpoint inhibitor in advanced-stage cancer patients.
In a prospective clinical study, miRNAs are isolated from the plasma of patients of various cancer types prior to immunotherapeutic treatment. Once reverse-transcribed into cDNA, expression profiling of miRNAs was performed using a multi-gene, amplification-based detection system that simultaneous analyzes over 150 miRNAs. Data were preprocessed, modeled and trained on 51 patients to derive a prediction algorithm, which was validated on 20 patients.
Three months after the administration of the therapy, patients were examined for response and categorized into two groups, disease control (CR/PR) or disease progression (PD) based on RECIST 1.1. A prediction algorithm was derived based on the miRNA expression patterning of 51 patients using data-driven approach. When validated on 20 patients, the overall accuracy of the response prediction is 80%, with a PPV of 73% and a NPV of 89%.
Utilizing data-driven modeling, circulating miRNA classifier shows significant predictive power. Validation is currently being performed in patients undergoing treatment. We believe this would be one of the first evidence demonstrating circulating miRNAs as potential predictive biomarkers for immunotherapy response in advanced-stage cancer patients.
Linkou Chang Gung Memorial Hospital.
Quark Biosciences, Inc.
C-H. Hsieh: Advisory / Consultancy: Quark Biosciences. S-T. Kang: Full / Part-time employment: Quark Biosciences. W-M. Chen: Full / Part-time employment: Quark Biosciences. Y-S. Hsieh: Full / Part-time employment: Quark Biosciences. E.P. Yang: Full / Part-time employment: Quark Biosciences. All other authors have declared no conflicts of interest.